Look, I'm going to say it again. Testing rate *per capita* is almost irrelevant because the spread rate is determined by the number of person to person contacts within a population, NOT the total number in the population in that area. IT'S MATHS, people.

And look. My husband literally used maths amongst other things to write a scheme that literally reduced the energy use in Australia in 2009. First class hons Maths & Phys, phD engineering physics. And he agrees with me.

So if the spread rate is going down it's because the person to person contacts are. And that's a no freaking brainer.

I'm going to break this down. Say you have two populations, one of 4,000, one of 40,000. Say you have a spread rate of doubling every 4 days. One person introduces the virus to each on day one.

Say 32 days pass before anyone realises the disease is in the population, so no social distancing. That's 8 cycles. In both populations, 256 people have the disease.

Say you aim for the per capita rate of 1 in 119 people (South Korean gold standard). In the population of 40,000, you only need 333 tests to get to that rate. *If* you've targetted extremely well, you could have caught all the carriers.

In the population of 4000, to beat the Korean rate of 1 in 119, you only need test 33 people. You've missed 223 infections. And this is why the **per capita* rate is not in itself a good enough measure of whether you are testing well enough.

You need to consider the spread rate, the targetting, and any spread mitigation policies. Now let's look at the 40,000 population with 333 tests. Let's assume a systematic bias to the testing - only testing known C19 contacts. Let's assume one case slips out at the 2nd cycle.

It was a missed contact. So of the 256 cases, 120 or so have come from the 2nd person infected, who wasn't traced. The testing regime is like Australia's - specifically exempting people who haven't got known exposure & who haven't come from overseas.

That's what's called a systematic bias in experimental design - a fault which will tend to return a result systematically missing something that exists in the field. In this hypothetical case, 120 mildly symptomatic or asymptomatic carriers.

These people won't show up until possibly a month after they are infected, when 5 will need ICU care. Now, say the government of this population promotes social distancing and heavy lockdown. It's possible that this slows the spread rate to a relative trickle. Maybe we are here.

Which is excellent! But people having been lulled into a false sense of security by the repeated citing of the per capita test rate let their guard down. They stop handwashing, all shop at once (supermarkets get crowded now because there's nothing else to do).

And then it's loose. But we just *do not know* until people start getting very sick. Or they don't.

So what's the solution?

Obvious. Broaden the testing criteria to see if it's out there. It doesn't even have to be much, just in suburbs where there have been outbreaks.

Just sample a few places outside the criteria to see how the efforts are going.

I'm banking on the lockdown measures we currently have saving us from getting to overseas measures. But I am pissed off with the patronising rubbish from people citing the per capita test rate without getting it into their heads that it's a misleading standard.

*BSc, experimental psychology, married to & tutored by a guy with double first class Maths & Phys, PhD engineering physics.